Exploration of new biomes to obtain microorganisms of industrial interest can reveal a number of new strains but they are not all equally efficient in their carotenoids production under different environmental conditions. Screening is needed to select the elite strain and thus commercially produce the pigment. The main objective of this study was to increase the carotenoid production capacity of a recently isolated yeast obtained from the Cerrado (Brazilian tropical savanna) biome. Four hundred and seventy yeast colonies were isolated, of which 25 were pigment-producing yeasts. The promising yeast was identified, by molecular analysis (polymerase chain reaction - PCR) as Rhodotorula lactose, presenting 4.0 g/L of biomass and 776.67 µg/L (194.17 µg/g) of carotenoids (volumetric and specific) in the commercial medium. In the optimization, the carbon source (43 g/L of glucose) was the most significant and positive factor in the carotenoid production process (1318.27 µg/L and 145.58 µg/g), followed by the concentration of nitrogen sources (0.7 g/L yeast extract and 0.7 g/L peptone). The equation generated from the optimized process was validated and showed similar values in the carotenoid production (1233.80 µg/L and 156.60 µg/g) proving it to be promising process model for a scale-up using submerged culture.